C3′/C4′-Stereochemical Effects of Digitoxigenin α-L-/α-D-Glycoside in Cancer Cytotoxicity

Full text of DOI:10.1002/cmdc.201200465

CHEMMEDCHEM
COMMUNICATIONS
DOI: 10.1002/cmdc.201200465
C3’/C4’-Stereochemical Effects of Digitoxigenin a-l-/a-d-
Glycoside in Cancer Cytotoxicity
[a]
[a]
[a]
[a]
John W. Hinds, Sean B. McKenna, Ehesan U. Sharif, Hua-Yu L. Wang,
[b]
[a]
Novruz G. Akhmedov,* and George A. O’Doherty*
Digitoxin, a potent cardiotonic agent that has been used con-
tinuously over centuries for the treatment of congestive heart
failure, has more recently been recognized for its potential ap-
moved to study the importance of the stereochemistry and
pattern of oxygenation (Figure 1). For example, inversion of
the C3’-hydroxy group of rhamno 1 gives altro 2, and removal
of the C3’-hydroxy group of altro 2 gives ascarylo 3. Similarly,
inversion followed by a removal of the C4’-hydroxy group of
ascarylo 3 gives C2’-epi-colito 5 and C3’-C4’-dideoxy-rhamno 6.
As a control, we also planned to prepare and test several d-
sugar diastereomers, which should have very similar pharma-
cokinetic properties but occupy drastically different three-di-
mensional space. Herein, we report the successful syntheses of
new digitoxin C3’/C4’-sugar analogues and the evaluation of
their cytotoxicity against four different cancer cell lines, includ-
ing lung cancer (NCI-H460 and A549) and breast cancer (MCF-
7 and MDA-MB-231).
[
1–3]
plication in oncology.
The mechanism of action for digitoxin
2
+
cardiotonic effect occurs via an increase of Ca
influx as
+
+
a result of inhibition of the extracellular Na /K -ATPase
[
4]
pump. In contrast to the cardiotonic effects, less detail is
[
5]
known for the mechanism of antitumor activity. It has been
shown that cell death occurs via apoptosis. Several studies led
[
6–8]
by Xie
have suggested that, in cancer cells, a cardiac glyco-
+
+
side-bound Na /K -ATPase complex signals via Src-tyrosine
kinase and 1,4,5-triphosphoinositol (IP3), several downstream
events that ultimately lead to apoptosis. This signaling path-
+
way is believed to occur through specific isoforms of the Na
+
/K -ATPase pump, which are associated with cancer and not
heart tissue.
These promising findings of potent cytotoxicity against
cancer cells inspired many structural–activity relationship (SAR)
studies, primarily aimed at understanding the role that the car-
bohydrate plays in the cytotoxicity of the cardiac glycosides.
These studies include direct comparison of glycosidic link-
[
9,10]
[10,11]
age,
the study of sugar-chain length,
the stereochemi-
[
12]
cal survey of digitoxin monosaccharides, and the study of
[
13]
C5’-alkyl steric effect. These carbohydrate-based SAR studies
have led to the discovery of new sugar motifs that significantly
[12]
improved the cytotoxicity across a range of cancer cell lines.
To date, we have found that digitoxin a-l-rhamno-monosac-
charide 1 to be the most active analogue across the widest
range of cell lines, with a-l-amiceto 4 having the next best re-
sults. This surprising similarity in anticancer activity of the C2’/
C3’-dideoxy-rhamno-analogue (a-l-amiceto 4) motivated us to
explore other deoxygenation patterns that could further en-
hance the cytotoxicity.
Using our lead digitoxin analogue (a-l-rhamno 1) as a start-
ing point, we conducted a systematic SAR study, focusing on
the stereochemistry and functionality of the C3’- and C4’-ste-
reocenters. Using de novo carbohydrate synthetic methodolo-
Figure 1. C3’/C4’-stereochemical study of digitoxin rhamno 1.
Retrosynthetically, synthesis of the desired C3’/C4’-sugar
congeners of digitoxin a-l-rhamno 1 can be prepared from
a common allylic alcohol intermediate (7) through a series of
oxidation and/or reduction sequences to install the required
hydroxy or hydride functional group (X=OH or H)
[
14,15]
gy,
the C3’- and C4’- hydroxy groups can be inverted or re-
[a] J. W. Hinds, S. B. McKenna, E. U. Sharif, Dr. H.-Y. L. Wang,
[16,17]
(
Scheme 1).
The stereochemically pure C2’-allylic alcohol
Prof. G. A. O’Doherty
Department of Chemistry & Chemical Biology, Northeastern University
could be delivered via a 1,3-allylic transposition of the epoxide
[15,18]
117 Hurtig Hall, 360 Huntington Ave., Boston, MA 02115 (USA)
ketone (8) using a Wharton rearrangement.
Epoxide
E-mail: g.odoherty@neu.edu
ketone intermediate 8, in turn, could be derived from a diaste-
reoselective epoxidation of the pyran enone, which could be
easily furnished by using a palladium-catalyzed glycosylation
to couple tert-butyloxycarbonyl (Boc)-protected pyranone 9 to
digitoxigenin (DigOH). The pyran sugar building blocks with
[
b] Dr. N. G. Akhmedov
Department of Chemistry, West Virginia University
2
17 Clark Hall, Prospect Street, Morgantown, WV 26506 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201200465.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 63

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
To install the desired ascarylo stereochemistry (i.e., 3-deoxy-
rhamno), we utilized a trans-diaxial epoxide opening approach
using a formal hydride source to establish the deoxygenation
[13]
pattern of the targeted sugars (3 and 17).
As shown in
Scheme 4, epoxide ketones a-l-8 and a-d-14 were first sub-
Scheme 1. Retrosynthetic analysis of digitoxin analogues.
absolute a-d- and a-l-sugar configuration can be prepared
from commercially available 2-acetylfuran 10 in three steps:
Scheme 4. Synthesis of a-l-/d-ascarylo-analgoues 3 and 17. Reagents and
[
19]
Noyori asymmetric reduction,
Achmatowicz oxidative rear-
conditions: a) NaBH
4
, CeCl
3
/MeOH, ꢀ788C, 2–4 h; b) MgBr
2 2
·Et O, AcOH, RT,
[
20]
rangement (N-bromosuccinimide (NBS)/H O), and tert-butox-
4 h; c) (Me Si) SiH, AIBN, toluene, reflux, 3 h.
2
3
3
yl-carbonylation, to give the desired pyranones (i.e., a-l-9 and
a-d-11)
with
good
a-diastereoselectivity
(a/b=4:1)
[
21]
(Scheme 2).
jected to Luche conditions (NaBH /CeCl ) to insert the desired
4 3
equatorial C4’-hydroxy group, and the resulting epoxides were
further opened regioselectively through a trans-diaxial path-
way under Lewis-acid-catalyzed conditions (MgBr ·Et O/AcOH)
2
2
to give bromides 15 and 16 in good yields. The bromide sub-
stituents were readily removed via radical-induced debromina-
tion (2,2’-azobisisobutyronitrile (AIBN)/(Me Si) SiH) to afford de-
3
3
Scheme 2. De novo synthesis of a-l-/a-d-Boc-pyranone. Reagents and condi-
tions: a) Noyori (R,R), HCO Na, 10 mol% CTAB, RT, 24 h; b) NBS, THF/H
3:1), NaHCO , NaOAc·3H O, 08C, 1 h; c) Boc O, DMAP, CH Cl
, ꢀ788C, 12 h;
d) Noyori (S,S), HCO Na, 10 mol% CTAB, RT, 24 h.
sired digitoxin a-l- and a-d-ascarylo analogues 3 and 17.
Based on our recent success with the incorporation of
a Wharton rearrangement into the synthesis of several unusual
2
2
O
(
3
2
2
2
2
2
[15]
and deoxy sugars, we decided to employ this new post-gly-
cosylation transformation to prepare the targeted digitoxin an-
alogues. Using optimized Wharton reaction conditions
(N H ·H O/AcOH (5:4) in methanol at 08C), a-l-8 and a-d-14
Our study began with the synthesis of a-l-/d-epoxide ke-
tones from digitoxins, a-l-aculo 12 and d-aculo 13 (Scheme 3).
2
4
2
were rearranged into allylic alcohols 7 and 18, respectively, in
moderate yields (Scheme 5). The resulting C3’=C4’ pyran
double bond was readily oxidized using osmium tetroxide/N-
methylmorpholine-N-oxide (NMO) (conditions A) or reduced
using 2-nitrobenzenesulfonyl hydrazine (o-NO PhSO NHNH )/N-
2
2
2
Scheme 3. Synthesis of a-l-/d-epoxide ketone. Reagents and conditions:
a) Pd
2 3 3 3 2 2 2 2
(dba) ·CHCl /PPh , DigOH, CH Cl , 08C, 6–8 h; b) H O , LiOH, MeOH,
0
8C, 45 min.
[
11]
As previously reported, Boc-protected pyranones 9 and 11
were readily coupled with digitoxigenin using palladium(0)-cat-
alyzed glycosylation to afford digitoxins, a-l-aculo 12 and d-
aculo 13, respectively, with complete retention of a-stereo-
chemistry. The enone moiety was further epoxidized with hy-
drogen peroxide and a catalytic amount (10%) of lithium hy-
droxide as a base to afford the desired a-l-/d-epoxide ketones
Scheme 5. Synthesis of altro-analogues 2, 6, 19 and 20. Reagents and condi-
tions: a) N ·H O, AcOH; b) OsO /NMO (aq), tBuOH/acetone (1:1), 08C, 6–
8 h; c) o-NO PhSO NHNH , Et N/NMM, RT, 1 d.
8
and 14 with excellent diastereoselectively (i.e., only one
2
H
4
2
4
isomer was observed).
2
2
2
3
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 64

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
methylmorpholine (NMM) (conditions B) to give the altro- or
(NCI-H460, A549, MCF-7, and MDA-MB-231) using a-l-rhamno
1 as a positive control and natural digitoxin as a standard to
study the importance of the C3’/C4’-sugar stereochemistry, as
well as the deoxygenation pattern. The method used to screen
these digitoxin analogues was optimized from our previously
reported MTT assay, where the compounds were tested at in-
C3’,C4’-dideoxy-rhamno stereochemistry in both l- (i.e., 2 and
6
) and d- (i.e., 19 and 20) diastereomeric analogues.
Using a similar approach toward the synthesis of 3-deoxy
sugars via epoxidation and trans-diaxial epoxide opening
Scheme 4), the C3’=C4’ olefin of Wharton-rearranged inter-
mediate 7 was treated with meta-chloroperoxybenzoic acid
mCPBA) to afford epoxy alcohol 21 as a single isomer with ex-
(
ꢀ
9
ꢀ5
creasing log concentrations from 10 to 10 m for 72 h in
complete growth medium containing 10% fetal bovine serum
(FBS) as opposed to a serum-free medium as used in our
(
cellent facial selectivity (Scheme 6). Epoxide 21 was then re-
duced regioselectively using magnesium bromide diethyl
etherate in acetic acid, followed by an AIBN-promoted
[12]
former procedure. This method provides a faster screening
protocol with higher accuracy in cytotoxicity evaluation. As
a result of the use of this new MTT protocol, it is important to
note that the IC50 values reported herein cannot be directly
compared with our previous data; however, the magnitude of
the changes in cytotoxic potency are comparable and in this
context, we used a-l-rhamno 1 as a control. The results for the
seven digitoxin a-l-analogues (1–6 and 15) and four corre-
sponding a-d-analogues (16, 17, 19 and 20) are given in
Table 1 and displayed in a cell-line specific manner in Figures 2
and 3.
(
Me Si) SiH debromination of the resulting axial bromide to
3 3
give desired a-l-2-epi-colito-analogue 5.
With all the desired digitoxin analogues in hand, we exam-
ined their cytotoxicity against four different cancer cell lines
While all the digitoxin analogues exhibited varying degrees
of cytotoxicity and sensitivity toward different cancer cells, a-l-
rhamno 1 and a-l-altro 2 were consistently the most active an-
alogues (Table 1). Surprisingly, C3’-stereoisomers 1 and 2
Scheme 6. Synthesis of a-l-C2’-epi-colito-analogue 5. Reagents and condi-
tions: a) mCPBA, NaHCO
3 2 2 2 2
, CH Cl , RT, 1 d; b) MgBr ·Et O, AcOH, RT, 4 h;
c) (Me Si) SiH, AIBN, toluene, reflux, 5 h.
3
3
Figure 2. Cytotoxicity of l-glycosides against cancer cells: A) NCI-H460; B) A549; C) MDA-MB-231; D) MCF-7.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 65

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
group in ascarylo 3 can be seen upon its removal to form the
C3’-C4’-dideoxy-rhamno 6 (i.e., approx. tenfold decrease in ac-
tivity from 3!6). In contrast, the axial hydroxy substitution at
C2’ is less pronounced. Thus, when the axial hydroxy group at
the C2’-position of ascarylo 3 was removed to form amiceto 4,
only small, albeit significant, effects were observed (approx.
twofold), and these effects occurred in both the positive and
negative direction.
Table 1. In vitro cytotoxicity of a-digitoxin analogues screened against
a panel of cancer cell lines.
[
a]
A convenient feature of this de novo approach to carbohy-
drate SAR studies is that the diastereomeric d-sugar analogues
can be prepared and tested. These d-sugar diastereomers ide-
ally should have nearly identical properties, such as solubility
and logP, to their l-sugar isomers. The cytotoxicity of these
diastereomeric l- and d-analogues against the four cell lines
was compared (Figure 3). Regardless of the cell line, direct
comparison of an l-sugar digitoxin analogue with its corre-
sponding d-sugar diastereomer consistently showed that the
[11]
l-sugar isomer is more active. This effect was most pro-
nounced for the most active altro-stereoisomers l-2 and d-19,
and could also be seen for the l-/d-diastereomeric pairs 3/17
and 15/16. In contrast, this l-/d-effect was too small to be ob-
served in the relatively inactive C3’-C4’-dideoxy rhamno-diaste-
reomeric pair l-6/d-20.
Compd
IC50 [nm]
MCF-7
NCI-H460
A549
MDA-MB-231
1
2
3
4
5
6
1
1
1
1
2
22ꢁ4
23ꢁ3
9ꢁ1
9ꢁ1
32ꢁ8
24ꢁ6
59ꢁ20
55ꢁ20
110 ꢁ20
53ꢁ10
10ꢁ4
50ꢁ20
58ꢁ20
460ꢁ200
ND
220ꢁ90
100ꢁ30
850ꢁ200
2900ꢁ2000
390ꢁ200
ND
In conclusion, we have demonstrated the utility of our
de novo approach for carbohydrate-based structure–activity re-
lationship studies. In this context, the Wharton reaction was
used for 1,3-allylic transposition in combination with other
post-glycosylation transformations. We successfully employed
this approach to prepare several novel carbohydrate motifs to
examine the importance of the C3’/C4’-stereocenters. We
found that the presence of a hydroxy group at the C4’-equato-
rial position is critical for potent cytotoxicity. The axial hydroxy
group at C2’ has a significant, albeit lesser, effect on activity;
whereas, a hydroxy at C3’ is substitutionally significant but not
stereochemically critical for cytotoxicity. This study helps to
define the limits of digitoxin monosaccharide carbohydrate
substitution on cytotoxicity and points to directions for further
analogue design. In contrast to substitution at the C-2,4 and 5-
positions, this study shows the potential for C-3 substitution
for Na/K-ATP-ase isoform selectivity without loss of activity.
Studies directed along these lines are ongoing, and the results
will be reported in due course.
22ꢁ4
290ꢁ70
1800ꢁ100
150ꢁ40
ND
150ꢁ30
140ꢁ60
24ꢁ4
5
6
7
9
0
110ꢁ50
32ꢁ8ND
360ꢁ200
170ꢁ100
ND
250ꢁ100
96ꢁ30
53ꢁ10
200ꢁ70
65ꢁ8
240ꢁ80
160ꢁ50
770ꢁ200
150ꢁ30
680ꢁ200
380ꢁ100
ND
270ꢁ90
DIG
110ꢁ50
[a] The IC50 value was measured based on the 95% confidence interval
using an MTT colorimetric assay (see the Experimental Section for details).
Data represent the mean ꢁSE of four independent experiments per-
formed in duplicate. ND: not determined (IC50 >10 mm). DIG: digitoxin.
showed nearly identical IC50 values across all four cell lines.
This is indicative of the importance of the C3’-hydroxy substi-
tution rather than that of the specific stereochemical orienta-
tion (i.e., inductive effects rather than covalent interactions of
the C3’-OH). This inductive effect can be seen in the C3’-deoxy
analogue (ascarylo 3), where a two- to fivefold decrease in ac-
tivity was found across the four cell lines. The three-dimension-
al limits of this spatial capacity can be seen in altro-bromide
Experimental Section
15. Thus, when the axial C3’-hydroxy group of 2 is replaced
Chemistry
with the larger, but similarly electron withdrawing, bromide,
around a fivefold drop in activity was observed across all cell
lines.
General: Air- and/or moisture-sensitive reactions were carried out
under Ar or N using oven-dried glassware and standard syringe/
2
Similarly, the presence of an equatorial C4’-hydroxy group
was also found to be critically important for activity. We found
that the stereochemistry of the C4’-hydroxy group has a much
greater impact on the cytotoxicity than the C3’-position. For
example, the activity dropped four- to eightfold across the
four cell lines when the equatorial C4’-hydroxy group of ascar-
ylo 3 was converted into the axial C4’-hydroxy group of C2’-
epi-colito 5. Once again, the importance of a C4’-hydroxy
septa techniques. Et O, THF, CH Cl and MeOH were dried by pass-
2 2 2
ing through an activated alumina column under Ar gas pressure.
Hexanes refer to the petroleum fraction bp=40–608C. Commercial
reagents were used without purification unless otherwise noted.
Flash chromatography was performed using the indicated solvent
system on silica gel standard grade 60 (230–400 mesh). Analytical
thin-layer chromatography (TLC) was performed with precoated
glass-backed plates (60, F254) and visualized by UV irradiation
(254 nm) or by staining with KMnO4 stain or anisaldehyde stain
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 66

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
Figure 3. Comparison of the effects of l- and d-diastereomers against cancer cells: A) NCI-H460; B) A549; C) MDA-MB-231; D) MCF-7. A student t-test was per-
formed to analyze each diastereomeric pairs of digitoxin analogues (N=8; ***, P<0.001; ****, P<0.0001).
2
D
2
ꢀ1
1
(
465 mL of 95% EtOH, 17 mL concd H SO , 5 mL AcOH, and 13 mL
[a] =ꢀ28.4 (c=0.38 mgdL , CH Cl ); H NMR (400 MHz, CDCl3):
2
4
2
2
anisaldehyde). R values were obtained by elution in the specified
solvent ratio (v/v). Specific rotations ([a]D ) were obtained using
a digital polarimeter in the solvent specified (concentrations are re-
d=5.88 (br, 1H), 4.98 (dd, J=18.4, 1.6 Hz, 1H), 4.86 (s, 1H), 4.80
(dd, J=18.0, 2.0 Hz, 1H), 4.05 (br, 1H), 3.98 (m, 1H), 3.92 (m, 1H),
3.76 (dq, J=9.6, 6.4 Hz, 1H), 3.56 (d, J=10.4 Hz, 1H), 3.48 (ddd, J=
10.4, 9.6, 3.6 Hz, 1H), 2.78 (dd, J=9.6, 6.0 Hz, 1H), 2.47 (d, J=
10.4 Hz, 1H), 2.21–2.07 (m, 2H), 1.92–1.40 (m, 17H), 1.34 (d, J=
f
temp
ꢀ1
1
13
ported as mgdL ). H and C NMR spectra were recorded on both
00 and 400 MHz spectrometers. Chemical shifts are reported rela-
6
1
13
tive to CDCl3 (d=7.26 ppm for H; for C d=77.23 ppm) or to
CD OD (d=4.81 ppm for H; d=49.15 ppm for CD OD for C).
6.0 Hz, 3H), 1.31–1.21 (m, 4H), 0.95 (s, 3H), 0.87 ppm (s, 3H);
C NMR (100 MHz, CDCl ): d=174.7, 174.6, 118.0, 97.7, 85.7, 73.6,
1
13
13
3
3
3
Multiplicities are reported using the following abbreviations: s (sin-
glet), d (doublet), t (triplet), q (quartet), m (multiplet) and br
72.9, 71.0, 69.8, 69.7, 65.6, 51.0, 49.8, 42.0, 40.1, 36.8, 35.9, 35.5,
33.3, 30.8, 29.2, 27.1, 26.7, 26.6, 24.0, 21.5, 21.4, 17.8, 16.0 ppm; IR
ꢀ1
(
broad). IR spectra were recorded on a Bruker FT-IR spectrometer;
(thin film): n˜ =3447, 2933, 2875, 1736, 1448, 1064, 1014, 735 cm ;
+
thin films were formed in CH Cl2 solution. Melting points (mp)
HRMS (ESI) m/z [M+Na] calcd for C H O : 543.29284, found:
29 44 8
2
were determined with electrothermal apparatus and are uncorrect-
ed.
543.29161.
(2S,3R,5R,6R)-2-Methyl-6-(digitoxigenoxy)-2H-pyran-3,5-diol (3):
(
2
0
2S,3R,4S,5R,6R)-3,4,5,6-Tetrahydro-2-methyl-6-(digitoxigenoxy)-
H-pyran-3,4,5-triol (2): A solution of allylic alcohol 7 (90 mg,
.185 mmol) in tBuOH/acetone (1.0 mL, 1:1 (v/v), 0.2m) at 08C was
A solution of 3-bromo-3,6-dideoxy-altrose 15 (96 mg, 0.164 mmol)
in anhyd toluene (1.65 mL) was treated with (Me Si) SiH (0.1 mL,
3
3
0.324 mmol) and solid AIBN (13.5 mg, 0.082 mmol) in one portion
at 08C. The inert atmosphere was secured by freeze/thaw vacuum-
ing followed by re-purging with Ar (ꢁ3). The mixture was then re-
fluxed at 808C for 3 h. The reaction mixture was cooled to RT and
directly loaded on silica gel for purification by flash chromatogra-
phy (70% EtOAc/hexanes) to give 3 as a white solid (56 mg,
0.111 mmol, 68% after chromatography), which was further puri-
fied by recrystallization from CHCl /hexanes. R =0.28 (80% EtOAc/
treated with a solution of NMO/H O (50% w/v, 0.20 mL). Crystalline
OsO (1.0 mg, 2 mol%) was added, and the reaction mixture was
stirred at 08C overnight. The reaction mixture was concentrated in
vacuo and directly loaded onto silica gel for purification by flash
chromatography (65!70% EtOAc/hexanes). Pure fractions were
combined and concentrated, and the resultant product was recrys-
2
4
tallized (CHCl /hexanes) to afford 2 as a white solid (85 mg,
0
3
3
f
2
D
3
ꢀ1
.163 mmol, 89%): R =0.57 (10% MeOH/EtOAc); mp: 134–1358C;
hexanes); mp: 224–2258C; [a] =ꢀ27.8 (c=1.0 mgdL , CH Cl );
f
2
2
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 67

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
1
H NMR (600 MHz, CDCl ): d=5.87 (br, 1H), 4.98 (dd, J=18.0,
medium and Dulbecco’s modified Eagle’s medium (DMEM) were
obtained from Mediatech Inc. (Manassas, VA, USA). Trypsin (0.25%
in EDTA) and MTT were purchased from Sigma–Aldrich (St. Louis,
MO, USA).
3
1
1
9
2
.2 Hz, 1H), 4.80 (dd, J=18.0, 1.8 Hz, 1H), 4.66 (s, 1H), 3.97 (br,
H), 3.83 (br, 1H), 3.69 (dq, J=9.0, 6.6 Hz, 1H), 3.59 (ddd, J=9.0,
.0, 4.2 Hz, 1H), 2.77 (dd, J=9.6, 6.0 Hz, 1H), 2.19–2.09 (m, 2H),
.06 (ddd, J=13.2, 3.6, 3.6 Hz, 1H), 1.91–1.82 (m, 4H), 1.75–1.36
Cell culture: NCI-H460 cells were cultured in RPMI-1640 medium
(
m, 17H), 1.26 (d, J=6.0 Hz, 3H), 1.26–1.20 (m, 2H), 0.93 (s, 3H),
ꢀ1
1
3
supplemented with 10% FBS, 2 mm l-glutamine, 100 UmL peni-
0
9
3
1
1
5
.87 ppm (s, 3H); C NMR (150 MHz, CDCl ): d=174.8, 174.7, 117.9,
3
ꢀ1
cillin, and 100 UmL streptomycin. A549, MCF-7, and MDA-MB-
7.2, 85.8, 73.7, 71.8, 70.0, 69.5, 68.5, 51.1, 49.8, 42.1, 40.3, 36.7,
5.9, 35.43, 35.42, 33.4, 30.6, 29.9, 27.1, 26.8, 26.7, 24.0, 21.6, 21.4,
7.9, 16.0 ppm; IR (thin film): n˜ =3436, 2934, 2879, 1738, 1046,
2
2
31 cells were cultured in DMEM supplemented with 10% FBS,
mm l-glutamine, 100 UmL penicillin, and 100 UmL strepto-
ꢀ
1
ꢀ1
ꢀ
1
+
mycin. Cell cultures were all maintained in a humidified atmos-
027, 732 cm ; HRMS (ESI) m/z [M+Na] calcd for C H O :
29
44
7
phere of 5% CO at 378C. Trypsin (0.25% EDTA solution) was used
2
27.29792, found 527.29862.
to detach the cells from the culture flask for plating and passaging.
(
2S,3S,5R,6R)-2-Methyl-6-(digitoxigenoxy)-2H-pyran-3,5-diol (5):
Method: NCI-H460, A549, MCF-7, and MDA-MB-231 cells were
seeded at a density of 5000 cells/well in a 96-well plate for 24 h
with complete growth medium. A dose-dependency study was run
for each sample with six log concentrations ranging from 10 mm to
A solution of digitoxigenyl 21-I (40 mg, 0.069 mmol) in anhyd tolu-
ene (1.50 mL) was treated with (Me Si) SiH (0.05 mL, 0.162 mmol)
and solid AIBN (5.6 mg, 0.034 mmol) as described above for com-
pound 3. The mixture was refluxed at 808C for 5 h before cooling
to RT for purification. Direct loading on to silica gel and flash chro-
matography (80% EtOAc/hexanes) gave 5 as a white solid (25 mg,
3
3
1
nm. Each concentration dose was prepared in serum-free
medium by a 100ꢁ dilution of the stock solution that was pre-
pared in DMSO. After seeding, cells were maintained in log-growth
phase for 24 h. The cells were then dosed with test compound so-
lution at the desired concentration (10 mL) and incubated at 378C
for 72 h. Cell viability was determined by incubating the cells with
0
.05 mmol, 72%), which was further purified by recrystallization
from CHCl /hexanes. R =0.40 (90% EtOAc/hexanes); mp: 136–
1
CDCl ): d=5.87 (s, 1H), 4.99 (dd, J=18.4, 1.6 Hz, 1H), 4.92 (s, 1H),
4
3
f
2
0
ꢀ1
1
378C; [a] =ꢀ26.5 (c=0.34 mgdL , CH Cl ); H NMR (400 MHz,
D
2
2
3
ꢀ1
MTT solution (10 mL/well, 4 mgmL ) in deionized water for 4 h.
.81 (dd, J=18.4, 1.6 Hz, 1H), 4.02 (ddd, J=10.8, 3.6, 3.2 Hz, 1H),
.94 (br, 1H), 3.90 (dq, J=6.4, 2.0 Hz, 1H), 3.71 (br, 1H), 2.78 (dd,
The medium was then aspirated, and the remaining formazan me-
tabolites were dissolved in DMSO (100 mL). The plate was then
shaken for 5 min, and the absorbance was read on a Gen5 Fluores-
cence Reader at 570 nm. For each concentration, sample were per-
formed in duplicate and four independent experiments were run
3
J=9.6, 6.0 Hz, 1H), 2.20–2.08 (m, 2H), 1.91–1.35 (m, 20H), 1.26–
1
.14 (m, 2H), 1.20 (d, J=5.6 Hz, 3H), 0.92 (s, 3H), 0.87 ppm (s, 3H);
1
3
C NMR (100 MHz, CDCl ): d=174.8 (2C), 117.9, 98.2, 85.8, 73.6,
3
7
3
1.7, 69.7, 66.0, 64.2, 51.1, 49.8, 42.0, 40.2, 36.8, 36.6, 35.8, 35.4,
3.4, 30.5, 29.6, 27.0, 26.8, 26.7, 24.0, 21.6, 21.3, 16.0 ppm; IR (thin
(N=8). Graphs were plotted and IC50 values calculated (reported in
ꢀ1
the 95% confidence interval) using GraphPad Prism version 5.0d
for Macs (GraphPad Software, San Diego, CA, USA).
film): n˜ =3440, 2928, 2857, 1737, 1087, 1056, 1026, 978, 751 cm
HRMS (ESI) m/z [M+Na] calcd for C H O : 527.29792, found:
;
+
29
44
7
5
27.29787.
(
2S,5R,6R)-5,6-Dihydro-2-methyl-6-(digitoxigenoxy)-2H-pyran-5-
Abbreviations
ol (6): A solution of allylic alcohol 7 (67 mg, 0.138 mmol) in NMM
0.5 mL) at 08C was treated with NBSH (150 mg, 0.690 mmol) and
(
Azobisisobutyronitrile (AIBN); N-bromosuccinimide (NBS); cetyltri-
methylammonium bromide (CTAB); meta-chloroperoxybenzoic acid
(mCPBA); dimethyl sulfoxide (DMSO); 3-(4,5-dimethylthiazol-2-yl)-
2,5-diphenyltetrazolium bromide (MTT); ethylenediaminetetraacetic
acid (EDTA); fetal bovine serum (FBS); N-methylmorpholine (NMM);
N-methylmorpholine-N-oxide (NMO); o-nitrobenzenesulfonyl hydra-
zine (NBSH); tetrahydrofuran (THF); tris(dibenzylideneacetone)di-
Et N (29 mg, 0.276 mmol). The resulting mixture was stirred with
gradual warming from 08C to RT over 24 h. The reaction mixture
was diluted with EtOAc (5 mL), quenched with 1n HCl solution
3
(
2 mL), and separated. The organic phase was washed with saturat-
ed aq NaHCO (3ꢁ2 mL), dried (Na SO ), filtered and concentrated
3
2
4
in vacuo. Purification by flash chromatography (50% EtOAc/hex-
anes) gave 6 as a white solid (61 mg, 0.124 mmol, 91%): R =0.45
palladium(0) (Pd (dba) ).
2 3
f
23
(
70% EtOAc/hexanes); mp: 127.8–1288C; [a] =ꢀ25.3 (c=
D
ꢀ1
1
0
4
1
2
.77 mgdL , CH Cl ); H NMR (400 MHz, CDCl ): d=5.87 (s, 1H),
2
2
3
.99 (dd, J=18.4, 1.6 Hz, 1H), 4.80 (dd, J=18.4, 1.2 Hz, 1H), 4.70 (s,
H), 3.97–3.89 (m, 2H), 3.57 (br, 1H), 2.78 (dd, J=9.6, 6.0 Hz, 1H),
.20–2.09 (m, 2H), 2.05–1.97 (m, 2H), 1.91–1.80 (m, 2H), 1.75–1.18
Acknowledgements
The authors are grateful to Prof. Yon Rojanasakul and Dr. Todd A.
Stueckle (both West Virginia University, Morgantown, USA) for
providing the cancer cell lines used in this study (NCI-H460, A549,
MCF-7, and MDA-MB-231) and for useful discussions. J.W.H. and
S.B.M. thank Northeastern University (Boston, USA) for a Provost
Undergraduate Team Research Award.
(
m, 21H), 1.14 (d, J=6.8 Hz, 3H), 0.93 (s, 3H), 0.87 ppm (s, 3H);
1
3
C NMR (100 MHz, CDCl ): d=174.7 (2C), 117.9, 98.0, 85.8, 73.6,
3
7
2
1.6, 66.5, 65.1, 51.1, 49.8, 42.1, 40.3, 36.7, 35.9, 35.4, 33.4, 30.7,
9.9, 27.3, 27.1, 26.8, 26.8, 25.7, 24.0, 21.7, 21.6, 21.4, 16.0 ppm; IR
ꢀ
1
(
thin film): n˜ =3446, 2929, 1741, 1448, 1117, 1030 cm ; HRMS (ESI)
+
m/z [M+Na] calcd for C H O : 511.30301, found: 511.30199.
29
44
6
Keywords: anticancer agents
palladium-catalyzed glycosylation · Wharton rearrangement
·
digitoxin analogues
·
Biology
Materials: Human lung epithelial (NCI-H460 and A549), breast carci-
noma (MCF-7), and breast adenocarcinoma (MDA-MB-231) cell
lines were generous gifts from Prof. Yon Rojanasakul (West Virginia
University, Morgantown, WV, USA) and were obtained from the
American Type Culture Collection (Manassas, VA, USA). RPMI-1640
[
1] S. Johansson, P. Lindholm, J. Gullbo, R. Larsson, L. Bohlin, P. Claeson, An-
ticancer Drugs 2001, 12, 475–483.
[2] M. Lꢂpez-Lꢃzaro, N. Pastor, S. S. Azrak, M. J. Ayuso, C. A. Austin, F.
Cortꢄs, J. Nat. Prod. 2005, 68, 1642–1645.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 68

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
[
3] R. A. Newman, P. Yang, A. D. Pawlus, K. I. Block, Mol. Interventions 2008,
[13] H.-Y. L. Wang, B. Wu, Q. Zhang, S.-W. Kang, Y. Rojanasakul, G. A. O’Doh-
erty, ACS Med. Chem. Lett. 2011, 2, 259–263.
8, 36–49.
[
[
[
[
4] W. H. Barry, Y. Hasin, T. W. Smith, Circ. Res. 1985, 56, 231–241.
5] M. Lꢂpez-Lꢃzaro, Expert Opin. Ther. Targets 2007, 11, 1043–1053.
6] M. Haas, A. Askari, Z. Xie, J. Biol. Chem. 2000, 275, 27832–27837.
7] Z. Yuan, T. Cai, J. Tian, A. V. Ivanov, D. R. Giovannucci, Z. Xie, Mol. Biol.
Cell 2005, 16, 4034–4045.
[14] M. Shan, Y. Xing, G. A. O’Doherty, J. Org. Chem. 2009, 74, 5961–5966.
[15] H.-Y. L. Wang, G. A. O’Doherty, Chem. Commun. 2011, 47, 10251–10253.
[16] H. Guo, G. A. O’Doherty, Org. Lett. 2005, 7, 3921–3924.
[17] M. H. Haukaas, G. A. O’Doherty, Org. Lett. 2002, 4, 1771–1774.
[18] P. S. Wharton, D. H. Bohlen, J. Org. Chem. 1961, 26, 3615–3616.
[19] A. Fujii, S. Hashiguchi, N. Uematsu, T. Ikariya, R. Noyori, J. Am. Chem.
Soc. 1996, 118, 2521–2522.
[
[
8] Z. Li, Z. Xie, Pfluegers Arch. Eur. J. Phys. 2009, 457, 635–644.
9] L. M. Langenhan, N. R. Peters, I. A. Guzei, M. Hoffmann, J. S. Thorson,
Proc. Natl. Acad. Sci. USA 2005, 102, 12305–12310.
[20] O. Achmatowicz, P. Bukowski, B. Szechner, Z. Zwierzchowska, A. Zamoj-
ski, Tetrahedron 1971, 27, 1973–1996.
[
10] A. K. V. Iyer, M. Zhou, N. Azad, H. Elbaz, L. Wang, D. K. Rogalsky, Y. Roja-
nasakul, G. A. O’Doherty, J. M. Langenhan, ACS Med. Chem. Lett. 2010, 1,
[21] R. S. Babu, M. Zhou, G. A. O’Doherty, J. Am. Chem. Soc. 2004, 126, 3428–
3429.
3
26–330.
11] H.-Y. L. Wang, Y. Rojanasakul, G. A. O’Doherty, ACS Med. Chem. Lett.
011, 2, 264–269.
[
[
2
12] H.-Y. L. Wang, W. Xin, M. Zhou, T. A. Stueckle, Y. Rojanasakul, G. A.
Received: October 10, 2012
O’Doherty, ACS Med. Chem. Lett. 2011, 2, 73–78.
Published online on November 8, 2012
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 63 – 69 69